Existing ruby maser oscillators employ three cavities, all held at the same very low temperature. This system is stabilized by a high-Q superconductor coated sapphire resonant cavity. Such oscillators are capable of frequency stabilities approaching 10.sup.-16, as characterized by the Allan variance of repeated frequency measurements. The Allan variance is found by repeatedly measuring the frequency, then finding the differences from previous measurements, and calculating the variance of the differences. These are then normalized by dividing by the oscillator frequency itself.
The ruby maser is the active electronic element in the oscillator, driving the three-cavity system into electromagnetic oscillation. The natural frequency of operation of the ruby maser is determined by the magnetic field existing at the ruby, biasing the energy levels of the chromium ions in the ruby. The bandwidth of the ruby is about 2% (corresponding to a Q of about 50), a value not sufficiently small enough to generate an ultra-high stability frequency without stabilization by a separate high-Q cavity. While the oscillation frequency is primarily determined by the high-Q stabilizing cavity, it also responds to the tuning of the maser. This takes place as follows:
Stabilized oscillation occurs at a frequency within bandwidths of both the ruby maser and the high-Q mode of the three-cavity electromagnetic resonator system. The actual oscillation frequency is determined primarily by that of the high-Q stabilizing resonator, but is also determined to a lesser extent by the maser. Specifically, if the maser's frequency varies by a certain amount, the oscillation frequency varies by this amount multiplied by the ratio of the maser Q (about 50) divided by the Q of the high-Q mode of the resonator (about 5.times.10.sup.8).
The magnetic field seen by the ruby is the sum of two components: the very stable external field applied by the superconducting magnet and "locked into" a superconducting cylinder surrounding the ruby, and an internal field caused by the polarization of the magnetic moments of the chromium ions. To make the ruby-biasing applied steady magnetic field be unvarying, the magnetic field is "locked into" a surrounding superconducting cylinder by causing the cylinder to pass into the superconducting state, by cooling it through the transition temperature, in the correct value of the applied magnetic field. This effectively isolates the ruby from external field changes, but not, of course, from its own magnetic polarization.
As the chromium ions change their occupation of the allowed energy states, the polarization field changes in magnitude. Each energy state corresponds to a different alignment of the magnetic moment of the chromium ion in the local magnetic field. Therefore, when a chromium ion changes its energy state, its contribution to the total polarization field also changes.
To produce the negative resistance that provides the amplifying action of the ruby maser, both a steady magnetic field and a radio frequency (RF) electromagnetic field (the RF pump signal) are applied to the ruby. When the frequency .nu. of the pump signal matches the separation of a pair of energy states, E.sub.i and E.sub.j, of a chromium ion, E.sub.j -E.sub.i =h.nu., where h is Planck's constant, the pump signal can cause the ion to undergo a transition to the other state. Under the proper circumstances, this action can be used to create an ion population in a higher energy level that is larger than the population in a lower level, thereby obtaining the capability to amplify a signal--as the high energy ions change downwards to the lower energy state, they release energy, thus acting as a negative resistance. But the action of stimulating transitions also changes the alignment of the ions' magnetic moments, and so changes the polarization field in the ruby.
Amplitude fluctuations in the pump signal leads to fluctuations in the polarization magnetic field in the ruby. This has been demonstrated in a device designed to detect low frequency modulation of a microwave signal by the use of a ruby maser; see I.A. Deryugin et al, "Signal Detection in a Maser", Radio Engineering and Electronic Physics, Vol. 17, pp. 270-271 (February 1972) (translated from Raadiotekhnika i Elektronika, Vol. 17, pp. 353-353 (February 1972)). In this experiment, the change in polarization field was sensed by means of a pick-up coil placed around the ruby maser and the low frequency electrical signal generated by the pick-up coil was then sensed by external electronics. In this way, the AC modulation of the microwave signal was detected.
Since the polarization field affects the operating frequency of the ruby maser, polarization fluctuations cause frequency fluctuations in the maser, and thus in the oscillator system. For a given circumstance of oscillator operation, a definite relation can be derived for the magnitude of oscillator frequency fluctuation resulting from the pump signal amplitude fluctuation of a specific size.
To obtain stable operation, this source of fluctuations must be reduced to acceptable levels. The common technique is to operate a power detector and an electronically variable attenuator in the signal transmission line. Variations in signal power are compensated for by responsively increasing or decreasing the attenuation. However, this technique cannot reduce the pump signal fluctuations sufficiently to allow oscillator stabilities better than about 10.sup.-16, using conventional power detectors and attenuators. Furthermore, these conventional components must operate at room temperature, whereas the amplitude that needs stabilization is that arriving at the ruby, at low temperature. Fluctuations in signal amplitude caused by changes in signal passage through the transmission line beyond the power detector will not be compensated.